appears to result from the humus occupying less space after than before solution. It may be reduced to an insignificant amount by the use of a large quantity of saturated borax solution. A certain amount of contraction is, however, observed when ferruginous clays free from humus (or freed from organic matter by long boiling with chromic acid) are shaken with strong soda. It is not exhibited by pure kaolin, and possibly may not be by soils containing little iron. In any case, it is complete in five minutes, whereas the contraction due to humus is proportional to the time of shaking. The very large contractions observed by Baumann seem to have resulted from the use of too small a quantity of borax solution, thus both enclosing far too much air in the decomposition flask, and at the same time failing to dilute the hypobromite solution sufficiently. Knop fills the flask to within 40 to 50 c.c., and uses much less hypobromite (10 c.c. to 100 grams) than Baumann seems to have done. To the objection that other nitrogenous substances give up nitrogen when treated with hypobromite, it is replied that from a manurial point of view they may be regarded as equal in value to ammonia, and considering the very small quantity of nitrogenous matter present, the fact that some of these substances do not give up the whole of their nitrogen can only introduce a most insignificant error. M. J. S. Estimation of Carbon in Arable Soils. By O. K. STACHOVSKY (J. Russ. Chem. Soc., 1887, 125-132). In 1882, Dokutschaéff's investigation "On Russian Black Earth'" was published, in which this author has shown the distribution and boundaries of arable soil and the "black earth" in Russia. Carbon was determined by Wolff's method, namely, oxidation with chromic acid. In 1883, Loges (Abstr., 1883, 247) investigated Danish soils and found that the results obtained by Wolff's method were too low as compared with those obtained by combustion with copper oxide. The same result had been obtained by Peak and Warington at an earlier period (Trans., 1880, 617). The author has investigated whether and how far the said difference refers to the case of the "black earth." He uses for oxidation not chromic mixture (the chromic sulphate formed interfering with complete oxidation--the lower numbers found by Loges may perhaps be due to this cause), but an aqueous solution of chromic anhydride. The carbonic anhydride is absorbed in Classen's apparatus (a vertical condenser and two soda-lime tubes). The results obtained by the author with different samples of "black earth" by the chromic acid method agree completely with those obtained by combustion with copper oxide, but of course the first method requires far less time than the second. When, however, substances containing humous matter, such as peat, were analysed, much lower results were obtained with chromic acid. The author thinks that the good results obtained with Russian “black earth" are probably due to the presence of large quantities of nitrogenous substances in "black earth," which by oxidation are possibly converted into nitric acid and nitrogen oxides; these being good oxygen carriers assist in completing the oxidation of the carbon. B. B. Estimation of Grape-sugar in Urine. By H. WILL (Arch. Pharm. [3], 25, 812-822).-The urine is mixed with an equal volume of one-fifth normal baryta solution, filtered, and to 5 c.c. of the filtrate another 5 c.c. of baryta solution is added, together with 100 c.c. of 90 per cent. alcohol; after agitation, the solution is set aside for two or three hours. The precipitate is collected, washed with 20 c.c. of 90 per cent. alcohol, and thrown, along with the filter, into the precipitating flask, in which 10 c.c. of decinormal sulphuric acid has previously been placed. After warming and shaking, the excess of sulphuric acid is neutralised with standard baryta solution, using as indicator not more than one drop of phenolphthaleïn solution (1 : 100). The barium sulphate is now filtered off, and the filtrate is evaporated to dryness and weighed. The grape-sugar remains as a yellowish mass containing a little baryta, which is obtained as carbonate on ignition. This weighing of the sugar serves as a check on the volumetric estimation by means of the baryta precipitate, as described. The author concludes that:-I. Grape-sugar in aqueous solution can be very accurately estimated by the baryta method, either by titrating the baryta or by weighing the separated sugar. II. When baryta in sufficient excess is present, the aqueous solution of the barium sugar compound is precipitated as BaO (C6H12O6)2 + BaO on the addition of so much alcohol that the whole mixture contains 81 to 86 per cent. of alcohol by volume. III. In the presence of only 68 to 70 per cent. of alcohol, the precipitate has the composition BaO(C6H12O6)2. IV. In diabetic urine, the results obtained with Fehling-Soxhlet's titration method agree closely with those yielded by the baryta method. J. T. Formation of Levulinic Acid, a Reaction for the Detection of Carbohydrates. By C. WEHMER and B. TOLLENS (Annalen, 243, 314-334).—Most of the contents of this article have already been published (Abstr., 1886, 532). The authors find that levulinic acid is not produced in appreciable quantities when normal urine is heated with mineral acids, showing that normal urine does not contain carbohydrates. W. C. W. Determination of Tartaric Acid. By F. GANTTER (Zeit. anal. Chem., 26, 714-719).-In carrying out Goldenberg's method (this vol., p. 327), Wiegert (Zeit. anal. Chem., 23, 359) adopts the following precautions. 1. The excess of potassium carbonate should be as small as possible. 2. Acetic acid must be added in corresponding excess. 3. It must be washed out long enough. 4. The acidified mixture must still contain water when the alcohol is added. To these the author adds that a larger quantity of substance than was originally prescribed must be used. Of argol, lees, or crude calcium tartrate, 10 grams, of residues, 30 grams should be taken, and in both cases one-half of the solution should be used for precipitation. The one or two minutes which is sufficient for the complete decomposition by acetic acid when working with pure substances, is not long enough with crude materials. The evaporation must be stopped at 20 to 30 c.c., the acetic acid added gradually and well stirred in, and the mixture warmed for 15 minutes on the water-bath. M. J. S. Determination of Tartaric Acid in Wine Lees and Tartar. II. By A. BORNTRÄGER (Zeit. anal. Chem., 26, 699-714).—In the experiments in Part I (Abstr., 1886, 1082), the precipitated potassium hydrogen tartrate was always allowed to deposit during the night. A series of comparisons is now made between the results obtained by stirring continuously for 10 minutes and filtering after half an hour, and those yielded by stirring for five minutes and then leaving during the night. The quantity of potassium hydrogen tartrate used was varied from 1.5 to 4 grams, and both 15 and 3 grams of oxalate was tried with each quantity of tartrate. Practically identical results (99-51 to 99-72 per cent.) were obtained in all cases. Grosjean's precaution to prepare the washing liquid exactly at the temperature of the air and to ascertain its actual saturation with potassium hydrogen tartrate by titration, is found to be needless. It suffices to take a hot solution of the tartrate, cool it to about 4° above the temperature of the air, add 10 per cent. of potassium chloride, stir 10 minutes, leave to stand for half an hour, and filter. Even if its temperature differs a few degrees from that of the air it is of no consequence. It should, however, always be freshly prepared. The following is a more convenient method than Grosjean's for determining the quantity of oxalate required for decomposing the calcium salts. The roughly weighed substance (7·5 grams of lees or calcium tartrate or 3.75 of tartar) is neutralised, and carbonic anhydride is boiled off. Oxalate is then added in quantities of 1.5 grams at a time, heating for 10 minutes after each addition, and testing a filtered sample (acidified with acetic acid) for oxalic acid. Since the amount ascertained in this way can never be more than 1.5 grams in excess, and since any amount between 15 and 3 grams is admissible, the amount so employed is increased by 15 grams in the actual analysis. In an artificial mixture made to represent lees, and containing calcium sulphate, an hour's heating with the oxalate was found still to leave some of the sulphate undecomposed. A precipitate of calcium oxalate was therefore obtained during the evaporation of the filtrate, and had to be filtered off. The paper concludes with a minute description of the author's mode of procedure, which, however, only differs from Grosjean's in so far as it embodies the suggestions of this and the previous communications. M. J. S. Qualitative and Quantitative Test for Resin Oils in Mineral Lubricating Oil. By L. STORCH (Chem. Centr., 1887, 1419; from Ber. österi. Ges. chem. Ind., 9, 93-95).—If to one or two drops of resin oil, 1 c.c. of acetic anhydride and a drop of concentrated sulphuric acid are added, a violet colour is immediately produced which soon changes to brown. The presence of fixed oils or mixtures of fats prevents the reaction. If the presence of resin oil is shown in a lubricant, the quantitative estimation is made as follows:-To a quantity of the oil, five times its volume of 96 per cent. alcohol is added, and the mixture shaken. The alcoholic solution, which contains the resin oil with a small admixture of the mineral oil, is poured off and evaporated. The residue is weighed and then treated with 10 times its volume of alcohol; any mineral oil undissolved is separated, and the alcohol again evaporated, and the residue weighed a second time. A correction for the amount of mineral oil left in the second residue can be made from the data furnished by the difference in weight of the two residues, and the amount of alcohol used. J. P. L. Modified Soxhlet's Apparatus. By J. J. BARLOW (Chem. News, 57, 56-57). In the apparatus described, the substance to be extracted, wrapped in a piece of calico, is suspended from a hook in the flask containing the ether. The flask is fitted with a double-bored cork, carrying two tubes, one short and bent to form a syphon whose longer limb delivers immediately above the substance, the other long and straight; these tubes are enclosed in a wider tube fitting over the cork, and connected with a condenser. The advantages claimed are simplicity, the constant exposure of the substance undergoing extraction to hot ether and hot ether vapour, and the easy recovery of the ether at the close of the extraction. D. A. L Determination of Butter in Milk. By H. N. MORSE and W. M. BURTON (Amer. Chem. J., 9, 222-231).-The following improvements are suggested for the process described (Abstr., 1887, 752). The mixture of milk and copper sulphate is ground whilst still somewhat moist, and in filling the extraction tube, the mixture is packed somewhat tightly by use of a glass rod. Treatment two or three times with light petroleum (15 c.c. each time) is ample for the extraction of the fat. The light petroleum solutions are heated on the water-bath for 20 minutes with so much potash dissolved in alcohol as will saponify 01 gram of butter, and finally the excess of alkali is estimated with hydrochloric acid, using phenolphthaleïn as indicator. Light petroleum, boiling at 45-60°, is as efficacious as that boiling at 30-45°; it is not necessary to remove the petroleum before saponifying; from the examination of a number of samples, it is shown that practically identical results are obtained by weighing the fat extracted, or by saponifying, or by conducting the whole operation as above, and that the amount of alkali required to saponify a given weight of butter is very constant; the variations noticed by Koettstorfer (0.2215 to 0.2324 gram potassium hydroxide per gram of butter) are due to the fact that when butter is fused and allowed to cool without agitation, a partial separation of the constituents takes place, and the mass is not homogeneous. The authors find that 0.02295 gram of potash is required for 0·1 gram of butter. H. B. Determination of the Dry Residue and Fat in Milk and Butter. By F. GANTTER (Zeit. anal. Chem., 26, 677–680).—As a porous material for absorbing the milk to be dried, the author prefers wood-fibre, such as is prepared for the manufacture of paper. It requires to be washed with light petroleum. About 2 grams are placed in a capsule, dried at 105°, and weighed together with a small glass rod. Being very hygroscopic, a cover is necessary. The milk (5 or 6 grams) is poured on the fibre, avoiding as far as possible any wetting of the capsule, and during the evaporation the fibre is stirred occasionally in such a way as to wipe the capsule clean. After one hour, it can be transferred to the drying oven, where 11⁄2 hour is long enough for drying the residue from 6 grams of even a very rich milk. The fibre is then removed from the capsule, wrapped in paper, and extracted in Soxhlet's apparatus with light petroleum. Here two hours' extraction is found to be sufficient. In determining the water in butter, 3 grams of fibre should be used for 5 grams of butter. Owing to the large surface exposed, a constant weight is obtained after 11⁄2 hours' drying. M. J. S. New Method of Examining Butter. By T. T. P. B. WARREN (Chem. News, 56, 262; compare this vol., p. 199).-Both cottonseed oil and cocoa-nut oil were found in a sample of oleomargarine, although the latter was free from taste and smell; subsequent experiment showed that by mixing cocoa-nut oil with a certain quantity of cotton-seed oil, its odour and taste are concealed; when, however, these two oils are present in oleomargarine, the animal fats separated from them have a strong, rancid odour. It is noteworthy that when cocoa-nut oil alone is treated with carbon bisulphide and sulphur chloride, it behaves like butter; but when mixed with most fats or oils it is not so easily acted on by sulphur chloride, unless in large excess; by applying the author's method (loc. cit.) to an experimental mixture, containing 1 part of cocoa-nut oil to 3 parts of cotton-seed oil, the former oil was recovered perfectly white and solid. example is given of the working of the method, applied to the analysis of an oleomargarine containing cotton-seed oil; and in two test experiments with the cotton-seed oil (as used in the analysis) alone, solid residues were obtained differing from one another by 5 per cent.; in the author's opinion, however, the difference need not be so great-after a little practice. D. A. L. An Action of Sulphur Chloride on Oils. By T. T. P. B. WARREN (Chem. News, 57, 26-27, 43).—The result produced by this reagent with linseed oil depends on the quality and quantity of the chloride, on the temperature at which the mixing is made, and on the quality and quantity of the solvent. If the solvent is volatile, inactive, and in excess, the reaction only starts when the sulphur chloride solution becomes sufficiently concentrated by the evaporation of the solvent. If the solvent is not volatile, the sulphur chloride reaction seems to be arrested or modified. Small quantities of sulphur chloride will merely thicken linseed oil owing to the altered oil dissolving in the unattacked oil, although oil once perfectly acted on will not dissolve in fresh oil. When analysing a mixture of drying with non-drying oils by means of sulphur chloride, it is well to add enough of a known oil to make the combined quantity of known and drying oil relatively large as compared with the non-drying oil. Under exactly similar conditions of experiment, a drying oil will yield constant quantities of a solid product insoluble in carbon disulphide; in the same way |